# Learning to Read Where to Look: Disease-Aware Vision-Language Pretraining for 3D CT

arXiv: https://arxiv.org/abs/2603.02026 · HTML: https://mlsanity.com/paper/2603.02026/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Medical vision-language models, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 15 of its 20 references that are in the map.
- First posted on arXiv 2 Mar 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Medical vision-language models](https://mlsanity.com/mllm/medical-vision-language-models/)

## Builds on (most cited first, 15 of 15)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features](https://mlsanity.com/paper/2502.14786/) (2025)
- [Generalist Foundation Models from a Multimodal Dataset for 3D Computed Tomography](https://mlsanity.com/paper/2403.17834/) (2024)
- [Unified Contrastive Learning in Image-Text-Label Space](https://mlsanity.com/paper/2204.03610/) (2022)
- [Merlin: A Computed Tomography Vision-Language Foundation Model and Dataset](https://mlsanity.com/paper/2406.06512/) (2024)
- [CXR-CLIP: Toward Large Scale Chest X-ray Language-Image Pre-training](https://mlsanity.com/paper/2310.13292/) (2023)
- [Machine-Learning-Based Multiple Abnormality Prediction with Large-Scale Chest Computed Tomography Volumes](https://mlsanity.com/paper/2002.04752/) (2020)
- [Large-scale and Fine-grained Vision-language Pre-training for Enhanced CT Image Understanding](https://mlsanity.com/paper/2501.14548/) (2025)
- [Pillar-0: A New Frontier for Radiology Foundation Models](https://mlsanity.com/paper/2511.17803/) (2025)
- [Bridged Semantic Alignment for Zero-shot 3D Medical Image Diagnosis](https://mlsanity.com/paper/2501.03565/) (2025)
- [Scaling Self-Supervised and Cross-Modal Pretraining for Volumetric CT Transformers](https://mlsanity.com/paper/2511.17209/) (2025)
- [Vision-Language Modeling in PET/CT for Visual Grounding of Positive Findings](https://mlsanity.com/paper/2502.00528/) (2025)
- [More performant and scalable: Rethinking contrastive vision-language pre-training of radiology in the LLM era](https://mlsanity.com/paper/2509.13175/) (2025)
- [MedVista3D: Vision-Language Modeling for Reducing Diagnostic Errors in 3D CT Disease Detection, Understanding and Reporting](https://mlsanity.com/paper/2509.03800/) (2025)

## Built on by (most cited first, 3 of 3)

- [Open-Ended CT Volume Segmentation with Weak Supervision from Language](https://mlsanity.com/paper/2607.25860/) (2026)
- [Volumetric Radiology AI in the Era of Multimodal Large Language Models](https://mlsanity.com/paper/2608.20549/) (2026)
- [FD-AA: A Lightweight Focal-Diffuse And Attenuation-Aware Head for Incidental Abdominal Abnormality Detection in Chest CT](https://mlsanity.com/paper/2609.36189/) (2026)

## Abstract (arXiv)

Recent 3D CT vision-language models align volumes with reports via contrastive pretraining, but typically rely on limited public data and provide only coarse global supervision. We train a 3D CT vision-language model on 98k report-volume pairs (50k patients) collected at a single hospital, combined with public datasets, using SigLIP-style contrastive pretraining together with prompt-based disease supervision in the shared vision-text embedding space. On CT-RATE, our model achieves state-of-the-art text-to-image retrieval (R@10 31.5 vs. 22.2) and competitive disease classification (AUC 83.8 vs. 83.8), with consistent results on Rad-ChestCT (AUC 77.0 vs. 77.3). We further observe that radiologists routinely reference specific images within their reports (e.g., ``series X, image Y''), linking textual descriptions to precise axial locations. We automatically mine 262k such snippet-slice pairs and introduce the task of intra-scan snippet localization -- predicting the axial depth referred to by a text snippet -- reducing mean absolute error to 36.3 mm at 12 mm feature resolution, compared with 67.0 mm for the best baseline. Adding this localization objective leaves retrieval and classification broadly unchanged within confidence bounds, yielding a single unified model for retrieval, classification, and intra-scan grounding.